no code implementations • 10 May 2023 • Jiawei Mao, Shujian Guo, Yuanqi Chang, Xuesong Yin, Binling Nie
During the fine-tuning phase, MSMAE is also driven by attention to the accurate masking of medical images.
no code implementations • 17 Mar 2023 • Jiawei Mao, Yuanqi Chang, Xuesong Yin, Binling Nie
Compared to other severe weather image restoration tasks, single image desnowing is a more challenging task.
1 code implementation • 28 Jan 2023 • Jiawei Mao, Rui Xu, Xuesong Yin, Yuanqi Chang, Binling Nie, Aibin Huang
POSTER achieves the state-of-the-art (SOTA) performance in FER by effectively combining facial landmark and image features through two-stream pyramid cross-fusion design.
Ranked #3 on Facial Expression Recognition (FER) on AffectNet
Facial Expression Recognition Facial Expression Recognition (FER)
1 code implementation • 12 Dec 2022 • Jiawei Mao, Honggu Zhou, Xuesong Yin, Yuanqi Chang. Binling Nie. Rui Xu
This results in ViT not performing as well as CNNs on small datasets like medicine and science.
1 code implementation • 24 Nov 2022 • Jiawei Mao, Guangyi Zhao, Yuanqi Chang, Xuesong Yin, Xiaogang Peng, Rui Xu
We extensively evaluate ASIT on facial datasets such as FFHQ and CelebA-HQ, showing that our approach achieves state-of-the-art facial inversion performance.
Facial Expression Recognition Facial Expression Recognition (FER) +1
no code implementations • 22 Nov 2022 • Honggu Zhou, Xiaogang Peng, Jiawei Mao, Zizhao Wu, Ming Zeng
To solve it, we proposed PointCMC, a novel cross-modal method to model multi-scale correspondences across modalities for self-supervised point cloud representation learning.
no code implementations • 11 Nov 2022 • Jiawei Mao, Yuanqi Chang, Xuesong Yin
The core mechanism of TT is the addition of a Class (CLS) token for summarizing window information in each local window.
no code implementations • 21 May 2022 • Jiawei Mao, Xuesong Yin, Yuanqi Chang, Honggu Zhou
The MIM paradigm enables the model to learn the main object features of the image by masking the input image and predicting the masked part by the unmasked part.
no code implementations • 29 Nov 2021 • Jiawei Mao, Xuesong Yin, Yuanqi Chang, Qi Huang
First, we combine with MixMatch to generate pseudo labels for the fake images and unlabeled images to do the classification.